【MFCC】MFCC增量用于计算相关性

算法思想:
$$设Maxtrix为N,J维矩阵、Size为M$$
$$Vecto{r}_1是表示Matrix的第1行$$
$$则：(NewVector{)}_i=\sum _{1\le j\le M}j\ast [(vector{)}_{(}i+j_{)}-(vector{)}_{(}i-j_{)}]$$
$$其中:if(i+j)>N:i+j=N；if(i-1)<0:i-j=0$$
$$i\in [0,N),i\in Z$$
代码实现

def calculate_delta(vector, correlation_numb=2):
    """
    Calculate and returns the delta of given feature vector matrix for evaluating correlation
    :param vector:
    :param correlation_numb:
    :return:
    """
    rows, cols = vector.shape
    deltas = zeros((rows, 20))  # initial vector of deltas
    n = correlation_numb   # 2 order vector shape
    for i in range(rows):
        index = []
        j = 1
        while j <= n:
            if i - j < 0:
                first = 0
            else:
                first = i - j
            if i + j > rows - 1:
                second = rows - 1
            else:
                second = i + j
            index.append((second, first))
            j += 1
        deltas[i] = sum([(i + 1) * (vector[index[i][0]] - vector[index[i][1]]) for i in range(len(index))]) / 10
    return deltas

具体例子

// 原矩阵Matrix
[[1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5]	Vector1
 [5 6 7 8 5 5 6 7 8 5 5 6 7 8 5 5 6 7 8 5]  Vector2
 [1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5]	...
 [5 6 7 8 5 5 6 7 8 5 5 6 7 8 5 5 6 7 8 5]	...
 [1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5]
 [5 6 7 8 5 5 6 7 8 5 5 6 7 8 5 5 6 7 8 5]
 [1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5]
 [5 6 7 8 5 5 6 7 8 5 5 6 7 8 5 5 6 7 8 5]
 [1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5]
 [5 6 7 8 5 5 6 7 8 5 5 6 7 8 5 5 6 7 8 5]]
 // 增量向量 size设置为3
 [(1, 0), (2, 0), (3, 0)]
[(2, 0), (3, 0), (4, 0)]
[(3, 1), (4, 0), (5, 0)]
[(4, 2), (5, 1), (6, 0)]
[(5, 3), (6, 2), (7, 1)]
[(6, 4), (7, 3), (8, 2)]
[(7, 5), (8, 4), (9, 3)]
[(8, 6), (9, 5), (9, 4)]
[(9, 7), (9, 6), (9, 5)]
[(9, 8), (9, 7), (9, 6)]
// 计算后的增量，
[[0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4 0.4 0. ]
 [0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8  0.8 0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0.  0. ]
 [0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8 0.8 0.  0.8 0.8 0.8  0.8 0. ]
 [0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4 0.4 0.  0.4 0.4 0.4  0.4 0. ]]